where is k_compat.h?

[James Ralston]

> A couple of other questions need to be answered before we can "go
> public" with this problem:
> 
> 1. If we back up to a 2.0.x series kernel, does the problem
>    disappear?
> 
> 2. Do other NICs suffer from the same full-duplex problem as 3COM
>    NICs?

I can't easily test #1, as I don't have a spare box at home to play
with, and I don't want to rebuild either of my two main machines.

There's also another question that needs to be answered:

3.  If we connected two machines together using a crossover cable,
    does the problem disappear?

If the answer is "yes", then this is clearly a duplex issue between
certain 3Com cards and certain switches.

> When throughput to my Linux box drops severely, performance is fine
> between Windows machines using the same 3com cards (in full duplex
> as well) on the switch.

I wouldn't put much faith in this as a metric, though, because I have
no idea how intelligent (or stupid) Windows is with respect to
managing full duplex NICs.  If Windows isn't smart enough to both read
from and write to the NIC's buffers at the same time (which wouldn't
surprise me), then it would be very unlikely that it would actually
cause the NIC to do anything in full duplex.

> I have another Linux box with kernel series 2.0.x setup for a
> customer and it is connected to a similar model 24-port LinkSys
> 10/100 switch.  There are never any performances issues there.  It
> also uses IP masquerade to the Internet, but its a dialup
> connection, rather than ADSL-to-Ethernet.  So this would seem to
> answer question #1 from above.

All the NICs are operating in full duplex, right?

This is sheer conjecture at this point, but if duplex issues are the
source of these problems, perhaps the reason why the 2.0 kernels don't
seem to be affected is because they weren't sufficiently threaded
enough to both read from and write to the NIC's buffers at the same
time.

I'll try to do my crossover cable testing within the next few days...

James

-------------------------------------------------------------------
To unsubscribe send a message body containing "unsubscribe"
to linux-vortex-bug-request@beowulf.org